This revised application is the continuation of research (R01 MH 48977; 8/93-8/96) showing that depletion of catecholamines via the tyrosine hydroxylase inhibitor alpha methyl para tyrosine or depletion of 5-HT through dietary depletion of the 5-HT precursor TRP can cause acute, transient return of clinical depression in some antidepressant-treated depressed patients. This suggests that synaptic levels of norepinephrine (NA) or 5-HT are necessary for the antidepressant response during the initial phases of treatment. Recent reports indicate that 5-HT reuptake inhibitor-treated depressed patients may be less likely to experience depressive relapse during TRP depletion the longer they have been maintained on medication. The present study will use TRP depletion as a probe to determine whether synaptic levels of 5-HT continue to be as important in later phases of treatment. The investigator hypothesizes that antidepressant responses involve a time-dependent process of adaptation that is only initiated by enhancement of 5-HT neurotransmission. However, as duration of treatment increases, antidepressant responses may be less vulnerable to disruption by transient depletion of 5-HT. The investigator will also compare pharmacotherapy and cognitive psychotherapy (CT) regarding the relative ability to maintain a therapeutic antidepressant response during acute 5-HT depletion over time. Method: Over a five-year period, 80 patients meeting DSM-IV criteria for current major depressive episode (MDE) will have baseline assessment and will then be randomly assigned to 6 months of treatment with either paroxetine (N=40) or CT (N=40). All subjects remaining in treatment will be tested at three times: after 4 weeks (N=80), 4 months (N=50) and 8 months (N=40). Paroxetine and CT will be continued during the first two testing times but both drug and psychotherapy will be discontinued 2 months prior to final testing. Testing at each time point includes two 3-day test sessions (TRP depletion and control) in a double-blind, placebo-controlled, counter-balanced crossover design. Each session involves a baseline day, a depletion day (or control), and a follow-up day. Behavioral ratings of mood and plasma TRP levels will be measured to assess response. Implications: Long-term maintenance of the therapeutic responses to antidepressant drugs may be more related to the adaptive changes in brain areas modulated by 5-HT than simply on restoration of synaptic levels of 5-HT. Investigating the role of 5-HT in the response to CT will allow inferences to be drawn about the similarities and differences in neurobiological mechanisms between pharmacological and psychosocial therapies.